Tony, How is this going? Have you been able to get CELT to run on a TI55xx? I am very interested in the % utilization of the hardware and whether multiple instance of CELT on the same DSP are feasible. I look forward to your findings. MikeH -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.xiph.org/pipermail/opus/attachments/20100321/fab36669/attachment-0002.htm
Hi, I never used CELT on the C55x myself, but from what I heard if set it to low complexity, then you can do encode+decode with around 60 MIPS on the C55x. Jean-Marc Selon Mike Hooper <mihooper at bellsouth.net>:> Tony, > > > > How is this going? Have you been able to get CELT to run on a TI55xx? I am > very interested in the % utilization of the hardware and whether multiple > instance of CELT on the same DSP are feasible. I look forward to your > findings. > > > > MikeH > >
Jean-Marc, Thank you for the feedback. My project does not require a bit rate as low as 64kbps. Do you have any thoughts on what effect allowing a higher data rate would have on the processor? I can use data rates as high as 300kbps. Will this allow the C55xx to use less MIPS, or process multiple instances of the algorithm simultaneously? If so, is there a simple mathematical formula (graph) of bit rate vs MIPS that you could provide? Thanks, MikeH -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.xiph.org/pipermail/opus/attachments/20100322/8f2bb75b/attachment-0002.htm